[SERVER-35041] Mobile SE Assert on system_errno 24 Created: 17/May/18 Updated: 06/Dec/22 Resolved: 22/Jul/19 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Storage |
| Affects Version/s: | 3.7.9 |
| Fix Version/s: | None |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Matt Lord (Inactive) | Assignee: | Backlog - Storage Engines Team |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | nonnyc, storage-engines | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Assigned Teams: |
Storage Engines
|
| Participants: |
| Description |
|
When getting a session to the database we check the status of the open call and blindly assert if it's not OK. A common reason for the status not being OK is the system error 24 (max open files limit) which results in CANTOPEN. We should handle this more gracefully and return an error up the call chain for the failed session creation rather than crashing the database instance. |
| Comments |
| Comment by Alexander Gorrod [ 18/May/18 ] |
|
sulabh.mahajan please take a look at this ticket? |
| Comment by Eric Milkie [ 17/May/18 ] |
|
Actually, I'm wondering if we need to audit a bunch more places that fassert and instead return errors, in the new world of being a library instead of a standalone process. For example, disk space exhaustion would be another place where we might consider not crashing and instead just returning an error (if possible). |
| Comment by Eric Milkie [ 17/May/18 ] |
|
How would a caller handle such an error? |